The third direction for this analysis is to use statistical analysis to develop categorisations for the data. Unlike above, there is no ground truth or researcher-driven rule settings. Explanatory factor analysis is an example of this research approach-and already widely used in social sciences. In explanatory factor analysis the aim is to reduce the number of variables by finding which variables relate to the same characteristics. The high number of variables is reduced to groups of variables, or formally, its dimensionality is decreased. The challenge is how to interpret the outcomes, as researchers must provide a meaning for the results.
This algorithmic data analysis approach is also known as unsupervised machine learning. Like supervised machine learning, there are many algorithms to do this analysis. However, there is no need for example data. Rather, the classification rules are determined only from data analysis.